Escolar Documentos
Profissional Documentos
Cultura Documentos
14alntenance,
June 1994
Spillway Gates
by
By.._
Distribution I
Availability Codes
I Avail andlor
Dist
Special
Final report
Ap1oved for public release; distribution is unllmited
Prepared for
Under
'
i.1
MWe following two leawer used as port of the number deuigneting tecnclasm reports of wasreatc published
under dhe Repair, Evaluation, Maintenance. and rehbilitaion (REMR) Reemah Ptpmrn identify the
pMblem ama under which the report was pmrep :
PrnhTem Area
CS
GT
HY
CO
Nobli
EM
El
OM
@PRINTED
ON RECYCLED PAPER
rea
Incude
biliorahicrefrenes
Wateways
E xprmntSain aawgn-nPulctlnDt
Strutrleautinolieeopiianae
b onE
oe
TA7~o
W34noREM-C-4
1.Hda
ulcgte.2 Silays
~mm -- Dttm
esloing-n
adontubcation -if
Waterways
M JhE..
aiteane
Evlain Experiment
evlatinalyrveedspillwayngineesig
.Structural Station. V ReataoiirPEvaluation,
/IyJh.
Bower,
.
Army.
Corps
of
Engineers.I1.USAryEgne
IIandrctrlehabltation
a.) ;Unit aed forates
Reeac Poram.teV.sTitle.yVI.tesnes: Technicalwe..
[ta];preportdfo(U.S. Army
EngineerWaewyExemntstaio)
xrsf
RE8-C-p i1 . 8c.Tcnclrpr
ERC-3
Waterways Experiment Station. IV. Repair, Evaluation, Maintenance,
and Rehabilitation Research Program. V. Title. VI. Series: Technical
report (U.S. Army Engineer Waterways Experiment Station);
REMR-CS-43.
TA7 W34 no.REMR-CS-43
Contents
Preface .....................................
vi
viii
ix
1- Introduction ................................
t ............
4
16
22
Introduction ................................
Lift Gates .................................
Tainter Gates ...............................
Roller Gates ................................
Summary of Findings ..........................
4-Task 3 - Environmental Effects on Gates .................
Damaging Structural Effects of Corrosion ................
Types and Identification of Corrosion Affecting Gates ........
Parameters Influencing Corrosion .....................
5-Task 4 - Effect of Repeated Loading on Gates ..............
Origin of Repeated Loads ........................
Fatigue of Riveted Structures ......................
Effect of Adding Weld Details to Riveted Structures .........
6-Task 5 - Structural Evaluation Guidelines .................
Components of a Structural Evaluation ..................
Application Examples ..........................
Recommendations for Continuing Evaluations ..............
22
25
27
31
31
33
33
34
39
43
43
44
51
54
54
69
73
tII
References ..................................
74
SF 298
Ust of Figures
Figure 1.
Figure 2.
Figure 3.
Figure 4.
Figure 5.
10
Figure 6.
13
Figure 7.
14
Figure 8.
23
24
Figure 9.
IV
24
25
26
26
Figure 14. Truss-framed tainter gate at Lock and Dam No. 2 .....
27
28
29
Figure 17. Submersible tainter gate at Lock and Dam No. 5 ......
30
31
Figure 19. Vertical lift gate at Bankhead Lock and Dam with paint
blistering and corrosion ....................
35
Figure 20. Tainter gate at Mississippi Lock and Dam No. 5 with
seal weld and localized corrosion ................
35
Figure 21. Vertical lift gate at Bankhead Lock and Dam with
crevice corrosion, corrosion at edges, and corrosion
on rivet heads .........................
36
41
Figure 23. Vertical lift gate at Bankhead Lock and Dam with
corrosion of rivet heads ....................
41
45
46
47
Figure 27. Fatigue crack from corrosion notch into rivet hole .....
47
48
49
50
51
52
53
59
59
Preface
The work described in this report wu sponsored by Headquarters, U.S.
Army Corps of Engineers (HQUSACE), as part of the Concrete and Steel
Structures Problem Area of the Repair, Evaluation, Maintenance, and Rehabilitation (REMR) Research Program and the Civil Works Guidance Update Program (CWOUP). The work was performed under the REMR
Work Unit 32641, "Evaluation and Repair of Hydraulic Steel Structures,"
for which Mr. Cameron P. Chasten, Information Technology Laboratory
(ITL), U.S. Army Engineer Waterways Experiment Station (WES), was
Principal Investigator and the CWGUP project "Structural Evaluation of
Existing Welded and Riveted Spillway Gates," for which Mr. Chasten also
was Principal Investigator. Mr. Don Dressier (CECW-EP) was the REMR
Technical Monitor, and Mr. William F. McCleese, Structures Laboratory
(SL), WES, was the REMR Program Manager. Mr. Paul Tan (CECW-ED)
was the CWGUP Technical Monitor for this work, and Mr. Thomas J.
Mudd, ITL, WES, was the CWGUP Program Manager.
Mr. William N. Rushing (CERD-C) was the REMR Coordinator at the
Directorate of Research and Development, HQUSACE; Mr. James E.
Crews (CECW-O) and Dr. Tony C. Liu (CECW-EG) served as the REMR
Overview Committee. Mr. James E. McDonald, SL, WES, was the REMR
Problem Area Leader.
The work was performed by the Center for Advanced Technology for
Large Structural Systems (ATLSS), Lehigh University, under U.S. Army
Corps of Engineers Contract Number DACW39-92-C-0063. The report
was prepared by Dr. John E. Bower, Mr. Mark R. Kaczinski, Mr.
Zouzhang Ma, Mr. Yi Zhou, Dr. John D. Wood, and Dr. Ben T. Yen,
ATLSS, under the general supervision of Mr. Chasten; Mr. H. Wayne
Jones, Chief, Scientific and Engineering Applications Center (S&EAC),
Computer-Aided Engineering Division (CAED), ITL; and Dr. N.
Radhakrishnan, Director, ITL. The report has been published by ATLSS
as "Structural Evaluation of Riveted Spillway Gates," ATLSS Report No.
92-12, Lehigh University, Bethlehem, PA, and much of the information
will be included in an Engineer Technical Letter, "Structural Inspection
and Evaluation of Existing Spillway Gates" (ETL 1110-2-351).
vi
Acknowledgment is expressed to Mr. L. E. Bridges, U.S. Army Engineer (USAE) District, Mobile, Operations Division, for arranging and assisting in onsite inspections at the John Hollis Bankhead Lock and Dam
on the Black Warrior River, and Messrs. Robert F. Post and Kent Hokens,
USAE District, St. Paul, Engineering Division, for arranging and assisting
in three onsite inspections at Mississippi River Lock and Dam Nos. 2, 5,
and 9. Acknowledgment is also expressed to the Structural Evaluation
Field Review Group (SEFRG), assembled to monitor this project, for providing assistance in this work. Members of the SEFRG and their affiliations are Messrs. Paul Tan, HQUSACE; Eugene Ardine, USAE Division,
Ohio River; Frank N. Johnson, USAE District, Vicksburg; David J. Smith,
USAE District, Omaha; and Thomas Sully, USAE District, St. Paul; and
Drs. John J. Jaeger, USAE District, Jacksonville; John E. Bower, ATLSS,
Lehigh University; and Chon L. Tsai, Ohio State University.
At the time of publication of this report, Director of WES was Dr. Robert W. Whalin. Commander was COL Bruce K. Howard, EN.
vii
Conversion Factors,
Non-SI to SI Units of Measurement
By
Fdeosi
ToObIln
dogmao
e0
0.304
moterI
hmiah(e
Oh)SWm4.W
meters
1.800347
llineis
14.5M0
-=ndper
ewape
kh
6094.757
v
"*
ixuemmmmmim
per meter
me
IeeC
m tw kolkig
Summary
The purpose of this study was to perform those tasks necessary to develop a set of guidelines that could be followed by Corps engineers to
achieve a structural evaluation of riveted gates. This has been addressed
through five tasks, identified herein as Tasks I to 5; a review of available
drawings, documents, and literature relative to riveted gates; an onsite inspection of gates at sites selected by the Corps; a study of how corrosion
might affect the gates; a study of loadings on the gates, especially repeated loadings; and, finally, the development of evaluation guidelines
using the findings from the other tasks and an example application of the
guidelines.
In Task I, one objective was to gain insight into the design principles
and materials that had been used for tainter, vertical lift, and roller type
gates in the 1930's when many gates were constructed, insofar as these
principles and materials affect evaluations and repairability today. Another objective was to determine what structural evaluations had been previously conducted and with what outcomes. It was determined that the
riveted gates had been designed as statically loaded structures; whereas,
more recent Corps documents reported that cyclic loads are occasionally
induced due to flow-induced vibrations caused by passing water. It was
also determined that buckling had occurred in gate members and that
some retrofit modifications had been made to curtail further buckling. A
review of steel standards suggested that the structural steels used in fabricating the riveted gates were generally not regarded as weldable steels,
which suggests that repairs by welding should not be indiscriminately
made on riveted gates.
In Task 2, ATLSS and Corps personnel made onsite inspections of gates
on the Black Warrior and Mississippi Rivers and observed tainter. vertical
lift, and roller gates. It was observed that a moderate amount of welding
had been done on most riveted gates. The inspections also provided direct
confirmation that vibrations were occasionally induced by the water flow;
however, it was observed that a slight adjustment of gate position (with
tainter gates) negated the vibration. Corrosion, accompanied by some degree of failure of the protective paint system, was prevalent on most gates.
Ix
In Task 3, the various forms of corrosion that can affect gates were
identified and assessed for their likelihood. The prevalent forms are depicted in photographs taken during the onsite inspections. The most frequently observed form of corrosion was crevice corrosion; although
pitting corrosion, galvanic corrosion, and general atmospheric corrosion
were also observed. In general, gate corrosion was accompanied by a failure of the paint system. Therefore, factors important to effective paint systems are described. Additional factors affecting corrosion at gate sites are
also described. Most of these are environmental: the pH of the river and
of rain, river water content including ions such as deicing salts, film-forming materials such as oil, and biological organisms. Corrosion has an important role in a structural evaluation because it can lead, depending on
circumstance, to loss of section, loss of strength, and diminished operability of the gate.
In Task 4, a general Corps concern about the effect of repeated loadings on riveted gates was examined. Although there may be several
sources of repeated loading, the most potentially damaging is probably the
flow-induced vibration that occurs when a tainter gate is open to some critical elevation (usually a few inches above the closed position). In the
study, the fatigue strength of riveted members is discussed and related to
standard categories of fatigue strength for welded details. It is concluded
that when the calculated (or preferably measured) nominal stress range at
a riveted detail does not exceed 6 ksi,l there is no concern for a fatigue
failure no matter what the age of the gate. (In this study, gate stresses
were not measured.) However, when the stress range is between 6 and
10 ksi, it is concluded that the fatigue strength of the riveted detail should
be taken equal to that for a Type C welded detail; and if the stress range
exceeds 10 ksi, the fatigue strength should be taken equal to that of the
more severe Type D welded detail. Inasmuch as both groove welds and filleted tack welds were observed during the onsite inspections, the fatigue
strength of these details should be equated to that of the even more severe
Type E welded detail.
In the final task, Task 5, a procedure for conducting a structural evaluation is given. Four components of an evaluation are discussed: preinspection assessment, inspection, assessment, and recommendations for
inspection, maintenance, and repair. For each component, the critical
question to be addressed by an engineer is presented, and the factors that
must be considered in his/her response are provided. The process is one
of steps, in that an assessment cannot be made until the pre-inspection assessment and inspection are made; and the final evaluation cannot be
made until all of the above are made. Included in Task 5 are discussions
of the critical areas to be inspected and of techniques for conducting inspections. The fatigue inspection guide is based on the broad experience
1 A table of factors for converting non-SI units of measure to SI is presented on page viii.
xi
Introduction
-odoctn
of limiting
Project Scope
In conducting this project, five tasks were incorporated into the
program:
a. A review of salient documents and drawings provided by the Corps
and/or available in the literature.
b. A site review of gates selected by Corps personnel.
c. A study of the environmental effects on gates, related to corrosion.
d. A study of repeated loadings on gates, and a study of the strength of
riveted structures under fatigue loading.
e. The development of evaluation guidelines using the results of the
aforementioned tasks.
The tasks and a brief description of each task are provided in the following paragraphs.
Camp w1 k m son
Task 3
Task 4
Corps personnel had suggested that fatigue considerations may be important for gates. Therefore, in this task, sources of repeated loadir., on
gates were reviewed and then fatigue data were applied to riveted gate details. A significant database on the fatigue behavior of riveted steel structural members exists, and much of these data, although derived mostly
from studies on bridge members, are transferrable to riveted spillway
gates. Some welded details were also examined when these details were
normally used with riveted gates, such as certain seal welds and tack
welds.
Task 5
In this task, the most compelling of the tasks, guidelines for a structural evaluation have been developed by identifying and discussing four
primary steps in an evaluation: (a) preinspection assessment, (b) inspection, (c) assessment, and (d) recommendations for inspection, maintenance, and repair. For each step, the primary question to be asked is
presented and alternative responses are discussed, leading to a judgment
on actions to be taken relative to the structural adequacy of the gate. Critical sites to be inspected for corrosion and fatigue damage are identified,
and inspection procedures are described.
Illustrative structural evaluations are also presented to demonstrate the
proposed guidelines. The examples, one actual and one assumed, are
based on one of the tainter gates on the Mississippi River. Finally, recommendations for continuing evaluations are presented.
Chaptr 1 Introduction
Task I - Review of
Literature
Literature on Gates
In this review, documents relative to tainter gates, vertical lift gates,
and roller gates are considered. In addition, documents describing some
special loadings, such as ice loading and dynamic loading, and structural
problems attributed to these loadings, are reviewed.
Tainter gates
In 1879 - 1881, Mr. J.B. Tainter and his partners developed improvements to the design of sluiceway gates and arrived at new designs
which would later carry his name and be called tainter gates. Those new
design ideas were patented (Tainter and Parker 1879; Parker et al. 1880;
Tainter 1881), and one of Tainter's gates is illustrated in Figure 1.
More than 60 years later, the Corps of Engineers had designed a simplified and lighter tainter gate (Figure 2), but one that retained many of
Tainter's original principles (Brizzel 1948). The major evolution during
this period included the following: (a) in material, steel was used instead
of timber, (b) in design, trunnion and hoist mechanisms were refined, and
(c) rivets and/or welding replaced other mechanical fasteners.
VericlloniuialScinyfGt
Figre
ecton. Ots
atetedby . BWTindlass188
Vericfal~, Longvitudial
SECTIONL
Secio
ELEVATION
ofGat
The tainter gate is often considered the most economical and usually
the most suitable type of gate for controlled spillways because of its simplicity, light weight, and low hoist-capacity requirements.
The principal elements of a typical tainter gate structure are the
following:
f. Seals. The side seal rubbing plates should be kept clean and smooth
to prevent corrosion and should be free from ice. There is movement between the bottom seal and its contact plate due to the live
load deflection of the ribs and the girders. For gates which are operated in subfreezing weather, it is necessary to deice the seals so that
gate operation can be maintained. The deicing is done by electric
heaters or air deicing devices.
g. Hoist. The chain or wire rope should contact the skin plate closely
for essentially the full height of the gate. The hoist cable is a nonredundant member; therefore, it should be kept in good condition
with little corrosion.
vertical lift gates
It is sometimes preferred to use vertical lift gates instead of tainter gates
in the following circumstances (U.S. Army Corps of Engineers 1962):
a. The elevation of the maximum controlled pool is so far above the sill
that excessively large piers would be required for tainter gates.
b. Flood discharges or drift conditions are such that any obstruction to
flow below the bottom of the spillway bridge is undesirable.
c. There is an overall economic advantage due to the speed of erection
of vertical lift gates and consequent shortening of the construction
time for the project as a whole.
Vertical lift gate assemblies are illustrated in Figures 3-5. Defined in
terms of how the water load is transferred to spillway piers, there are
three types of gates:
a. Fixed-wheel gate (Figures3 and 4). The wheels revolve on fixed
axles mounted on the end frame of the gate. This is the most common lift gate and is adaptable to long spans and a heavy moving
load. This type of gate was observed on the Black Warrior River
(see Chapter 3).
I-
ii
itl
14s
Ih~
2Tl -R~wo
I fw
jow
F.
W-
ItI
th of
I'CF
A*FUV
lplp
***'
Ic
E-
C~wp
LNWt~U
Tsk
1-
RGWW
S4w
10.
h..
10
e. Lifting arrangement. The lifting hooks attached to the gate and their
connections to the gate are of primary importance, since their failure would cause the gate to be inoperable. The lifting arrangement
must always be well maintained.
ure 7).
The primary structural members of a roller gate are the following:
a. Drum assembly. The drum assembly is a large cylinder which acts as
a beam and torque tube to carry the hydrostatic and dead loads. The
skin of the cylinder is stiffened by equally spaced stiffeners or ribs.
The stiffeners are braced at intermediate points along the length of
the drum by a truss-type assembly. The hydrostatic and dead loads
are transferred to the end disks.
b. Apron assembly. The apron assembly is an extension of the damming
surface. It consists of a skin plate extending outward from the cylinder and supported by horizontal ribs and a truss bracing assembly.
The connections between the drum assembly and the apron assemnbly are the transition area which needs to be inspected carefully.
c. End disks. The skin plate and horizontal stiffeners transfer load into
the end disks which are essentially truss-type configurations. The
loads are then transferred from the end disks to the lifting chains
and the piers.
12
Figure 6.
13
We.."
,.~*0-
7 J
.0:.
AL.
'
,**4
I/
Figure 7.
14
of ice loads on the gates and piers. The Corps of Engineers' Cold Regions
Research and Engineering Laboratory (CRREL) was also consulted. The
studies concluded that (USAED, St. Paul 1988):
a. The piers resist most of the ice loading when the ice is thick and/or is
fairly strong.
b. The design ice loading of 5,000 lb/ft of width is reasonable only
when the ice is not very strong, is on the order of I ft thick, and
yields completely around the piers.
c. The piers of the Mississippi River dams might cause a larger reduction in the ice forces than suggested by EM 1110-2-2702.
d. The 5,000-lb/ft ice loading is a guideline for design, not a
specification.
Flow-induced vibration. In the 1960's, vibration of tainter gates occurred (USACE 1971) at several dams on the Arkansas River. Some vibration was so severe that fatigue failures occurred in a number of the
structural members and welded connections of the gates. Later (USAED,
St. Paul 1986), tainter gates at Lock and Dam No. 5 on the upper Mississippi River experienced vibrations similar to those observed on dams on
the Arkansas River. Fatigue damage due to gate vibration was suggested
as possibly being one reason leading to the brittle failure of a trunnion
girder on a tainter gate at lower St. Anthony Falls Dam in St. Paul in 1981
(USAED, St. Paul 1986). Reportedly, vibration has also led to cracking in
the end shields of roller gates.]
The Little Rock and St. Paul Districts have conducted separate investigations of this problem. They monitored tainter gates with the problem of
vibration, made dynamic measurements in the field and performed a
model test in the laboratory, pursued intensive analytical work and developed some modified gate seals, and worked out a new gate operating pro-
force and debris impact could cause vibration, spillway gate vibration is
generally flow-induced vibration generated by water flow between the
gate bottom and sill. Flow-induced vibration primarily depends on interaction between the fluid and the gate structure. The factors which are related to flow-induced vibration are the following:
a. The flow velocity which is related to pool differential and gate
opening.
b. The fluid density.
PeronalCof=nam
15
tion will continue. The problems with operating gates in this way include
downstream scour damage due to locally high discharge velocities and a
more complicated operation.
The configuration of the gate seal is a major factor in setting up flow
conditions which cause vibration. Several gate seal modifications were developed by the Little Rock District (USACE 1971). After both model and
field tests, it was noted that it is necessary to provide a sharp break point
for flows at all gate openings and to stiffen the bottom cantilevered portion of the gate skin plate.
From the field measurements made on Gate 24 at Lock and Dam No. 5
on the Mississippi River, it was noted that the vibration frequency was
about 10.5 Hz and the relevant stress range was about 3.7 ksi on the girder
and about 4.4 ksi on the strut arm (USAED, St. Paul 1986).
Although it has been realized that flow-induced vibration can usually
be avoided by adjusting the operating procedure and employing proper
gate seals, the problem does occur and is of concern. Further study of this
phenomenon is still necessary.
I U.S. Army Corps of Engineers Upper Mississippi Valley Division. (1937). "Mississippi River
Lock & Dmn No. 25 Dam 60 ftx 25 ft Tainter Gate," Drawings M-L25-48W0A to 8.1.
2 U.S. Army Corps of Engineers Drawings. (1933-1937). "Mississippi River Lock & Dam No.
4," (a) General: M-L4-O/2-FS and 40/I to 3-FS; (b) Tainter gates: 40/55- and 56-FS and -48/I- to
15-FS; (c) Roller gates: 47/A- and B- and I- to 22-FS.
16
1
Tuscaloosa office of the Mobile District furnished drawings .2 of a verti-
cal lift gate at Lock & Dam No. 17 (currently named John Hollis
Bankhead Lock & Dam) on the Black Warrior River. These latter drawings were made in 1935.
The drawings of the Mississippi River gates indicated very little about
the steels used to fabricate the gates. "Structural steel" was the common
notation and apparently included structural steel plates, shapes (such as angles and channels), and rivets. The drawings of the Black Warrior River
gates used a different notation. Structural members were either "carbon
steel" or "silicon steel," depending on the size of the gate and application
of the member. Higher strength silicon steel (21,000 psi allowable stress
in tension and bending 2 ) seemed to have been used for main structural
plates and angles in 52-ft-wide gates whereas carbon steel (14,000 psi allowable stress') was apparently used for stiffeners and other nonstructural
elements on the 52-ft-wide gates and all components of the smaller 24-ftwide gates. The rivet steel was identified on the drawings as being carbon
steel in all cases. No specific references to ASTM designations were
found on these early drawings.
' U.S. AMy Cors of Engineers. (1935). 'ock &Dam No. 17, Black Warrior River, Ala.,
Crest Gates," Details of 52-ft Gate, Sheets No. I to 3. Mobile, AL.
17
steel, had a limit on sulfur content. A7 steel also was restricted to openhearth or electric furnace production, and excluded the older acid-bessemer production. These compositional and production restrictions suggest
that A7 bridge steel was recognized as the premium steel of the two.
For a brief period (1932-33), "structural steel" also could have been
supplied as ASTM A140 steel, which was a tentative replacement for both
A7 and A9 steels (AISC 1953).
"Silicon steel" as identified on the Black Warrior River drawings was
probably ASTM A94 structural silicon steel (ASTM 1925). This was a
high-strength steel with a specified minimum silicon content that attained
its high strength (minimum yield point of 45,000 psi and tensile strength
of 80,000 to 95,000 psi) through a high level of carbon (0.44 percent maximum). It also had limits on its phosphorus and sulfur contents.
An important characteristic of the early steels, regardless of whether
they were A7, A9, A140, or A94 silicon steel, is that they either had no
specified level or a high level of carbon in their composition. Consequently, the carbon level was either not rigorously controlled or was moderately high, with the result that the steels probably only had and have
poor to fair weldability. The specification for A94 structural silicon steel
specifically limits welding and specifies a preheat condition when welding
must be done. Of course, the steels were being used for riveted structures,
so weldability was not then a concern to designers. But it needs to be considered for weld repairs or maintenance contemplated today.
Up to date. In 1939, A7 and A9 were consolidated into a single specification, A7 steel for bridges and buildings (ASTM 1939), which then became the single specification for "structural steel." In 1954 a new
"structural steel for welding," A373 steel, was introduced (ASTM 1958).
Both A7 and A373 steels were consolidated in 1965 into the one specification, A36 steel (ASTM 1960), which is the basic structural steel today,
and is used for both welded and bolted applications.
Allowable and yield stresses. During the same period that A7 steel
was evolving, AISC changed its basic allowable working stress for
structural steel only once and raised it in 1936 from 18,000 psi to 20,000
psi (AISC 1953). The ASTM requirement for minimum yield point during
this period was generally 0.5 x tensile strength or not less than 30,000 psi;
in 1933, the minimum of 30,000 psi was raised to 33,000 psi for plate and
shape products. When A373 steel was introduced, that steel had a minimum yield point of 32,000 psi, suggesting that to improve weldability at
that time, some sacrifice in strength was necessary. Only when A36 steel
was introduced in 1960 in a tentative specification (ASTM 1960), did the
minimum yield point for structural steel plates and shapes increase to
36,000 psi. By that time, weldability and welding practices for structural
steel had markedly improved and standardized.
18
19
20
ChaPt
111-
IN
co
-t-A
-~
to-.
*l.~
41Q.100
Sc
ideA
"
ouh
lmUm
'Ar
Wl4m
Min
CV00.m
b-wp
ms
;m
2401
wImwSvmfl flStluIperbsubcS
NaPwknumbIwo
em4minbIguMa
ChqOFs
2 Tauko1-
qWsm
INS-
-m
Awll
*41111
l.
--
OaSorEa
alsmmmmof
moro2
Introduction
The locations selected by the Corps of Engineers for these inspection
visits included the John Hollis Bankhead Lock & Dam on the Black Warrior River in Alabama and Lock and Dam Nos. 2, 5, and 9 along the Upper
Mississippi River.
The John Hollis Bankhead Lock and Dam (Figure 8) is located in the
Corps' Mobile District, approximately 40 miles northeast of Tuscaloosa,
AL, and was visited by ATLSS personnel and Corps representative Mr.
L.E. Bridges (Mobile District - Tuscaloosa Office) on April 23, 1991.
The dam was originally built in 1915 and was modified in 1936 to increase the pool elevation by installing 22 fixed-wheel, single-section vertical lift spillway gates of riveted construction. Currently, 21 of the
spillway Sates are the original riveted structures, and the 22nd is a welded
gate installed in 1991. The reasons for this gate replacement are explained below. All Sates were reported to be operational and to maintain a
maximum pool differential of 69 ft at the dam with no tailwater conditions. The Holt Lake Lock and Dam, which is located downstream of
Bankhead Dam, was also visited. Because the spillway Sates at this facility are of welded construction, a discussion of the findings will not be included in this report.
i
22
K.R.
Kw
mpou of trip io Tuscdoos office, memo w prjec file, May 19, 1992.
ChWW~ 3 Tk
Figure 8.
Lock and Dam Nos. 2, 5, and 9 are all located on the Mississippi River
(St. Paul District) and were visited on May 27-28, 1992 by a team of
ATLSS and Corps personnel.' Dam No. 2 is located at Hastings, MN (Figure 9), and was originally built in the 1920's to maintain a pool differential of approximately 12 ft. The dam consists of 19 riveted tainter gates
constructed in a unique three-dimensional space truss configuration. Lock
and Dam No. 5 near Winona, MN (Figure 10), was built from 1933-1935
and consists of 6 riveted roller gates and 28 riveted tainter gates. A pool
differential of approximately 8 ft is maintained at the dam. Finally, Dam
No. 9 near Lynxville, WI (Figure 11), was also inspected. This structure
consists of five riveted roller gates and eight riveted tainter gates which
were built from 1934-1938 and are used to maintain a pool differential of
approximately 8 ft. The spillway gates at all three locations were reported
to be in operating condition. In fact, many of the gates are operated several times each week to control the river flow. Because high tailwater
conditions exist at these sites, the gates tend to accumulate a significant
amount of debris. No routine maintenance program is followed to remove
debris or touch-up painted areas. A discussion of the structural steel used
in construction of the spillway gates was presented in Chapter 2.
Observations of the riveted lift, tainter, and roller type spillway gates
visited at each site are summarized below.
M. R. Maczinski, report of trip to St. Paul District office. miemo to project file, June 17, 1992.
23
Figure 9.
24
Lift Gates
At Bankhead Lock and Dam, the built-up riveted gates are 52 ft,
3-3/4-in. long; 13 ft, 6-in. high; and 2 ft, 7-3/4-in. in depth (Figure 12).
Five cover-plated and stiffened girders are the main load and carry members of the gate and support a 3/8-in.-thick skin plate. A 3-in.-thick steel
casting, which acts as a base seal, is bolted to the bottom girder along the
entire length of the gate. A copy of a portion of the framing plan is attached. All structural elements (beam web, flanges, etc.) of the gate are
specified as "silicon" steel with an allowable tensile stress of 21 ksi. Stiffeners and other nonstructural elements are specified as "carbon" steel
with an allowable tensile stress of 14 ksi.
A thorough inspection was made of the riveted lift gate which was removed from the dam in 1991 and replaced by the welded gate. When inspected, this gate was lying down in a horizontal position rather than the
normal operating vertical position. No sign of structural distress (i.e., fatigue cracks, fractured or buckled members) or repaired members were observed, nor were any structural problems reported by Mr. Bridges. A
moderate amount of corrosion was evident on the top side of each girder
web due to the buildup of debris. As shown in Figure 13, more significant
corrosion was seen at the lower corners of the lift gate with some rivet
heads suffering 100 percent section loss. The cast steel base seal was also
corroded and irregularly eroded from the 56 years of service. This eroded
base seal did not make a watertight fit and was one of the reasons given
for replacing the gate. Another reason cited was the corrosion of the
lower corners of the gate; however, the most compelling reason for
Chapter 3 Task 2 - Site Review of Gates
25
16A
Figure 13. Heavy corrosion in lower corner of lift gate at Bankhead Damn
26
replacing the riveted gate was mechanical failure of the bearings in the
two reaction and guide wheels along each side of the gate. Mr. Bridges reported that, as a preventative measure, the 21 remaining riveted lift gates
will be replaced with welded structures with new reaction and guide
wheels as funds become available.
An Alabama Power Company hydroelectric generating plant is located
at the dam; therefore, water conservation is an important consideration in
gate operations. It was reported that the gates are only opened 3 to 4
times each year during very severe rain storms. During these periods the
gates are usually left open approximately 3 to 4 days.
Typically, one general inspection of a gate structure is made annually
as debris is cleaned from the gate and corroded areas are painted. According to Mr. Bridges, beyond this cleaning and inspection, no other program
is followed for inspecting the gates, and the lockmen are solely responsible for reporting any unusual behavior in the performance of the gates.
Tainter Gates
At Lack and Dam No. 2 each gate is 30 ft wide, 20 ft high, and 28 ft in
radius from the center of the trunnion bearing to the skin plate and is constructed as a three-dimensional space truss as shown in Figure 14. This
structural configuration is unique for tainter gates and consists of a series
of I I frames along the 30 ft width to carry the loads from the skin plate.
27
girder to the top of the gate. All nonsubmersible gates use Type J side and
bottom seal details which have been reported (USACE 1971, 1986) to be
prone to vibration problems. Corrosion was seen on some of the rivet
heads and along the top surface of the web on the upper horizontal girder
under the diversion plate.
Web and flange buckling on the strut arms adjacent to the knee brace
intersection from the upper horizontal girder was visible on several gates
and is most severe on gate No. 24. As reported in Corps documents, this
damage is believed to be a result of excessive ice loads on the structure
(Tante 1990, Hokens 1989').
With the assistance of lock personnel, Gate No. 23 was fully closed and
then reopened approximately 0.1 ft when vibration began normal to the
face of the gate. By rough measurement, the vibration frequency was
estimated at 5 to 10 Hz. The amplitude of the vibration was maximum at
midspan of the gate and was sufficient to create an audible noise and
make ripples in the backwater as shown in Figure 16. Although the vibration damped out towards the strut arms, it was still noticeable at the trunnion pins. It was reported that the gates would normally not be set in a
position which starts vibrating. No fatigue cracks were detected on the
structure. Although the gate is of riveted construction, groove welding
was used to water-seal the gaps between adjacent skin plates, and
IK.Hokem, Tainter gate strut arm bending at Lock and Dam No.5, memorandum for record
Figure 16. Effects of vibration at gate No. 23 on Lock and Dam No. 5
Chopter 3 Task 2 - Site Review of Gates
29
30
of Gati
Roller Gates
At Lock and Dam No. 5. the gates are 60 ft wide, 15 ft in diameter, and
20 ft in height when the roller apron is included. No structural distress,
significant corrosion, or paint blistering was visible. The rivet pattern on
these gates has a more uniform pitch and fewer transverse rows of rivets
than the roller gates at Dam No. 9.
At Lock and Dam No. 9, the gates shown in Figure 18 are 80 ft wide,
15 ft in diameter, and 20 ft in height when the roller apron is included.
No signs of structural distress were visible; however, the exposed surfaces
of the roller cylinder have developed excessive paint blistering. Pitting
corrosion and a small amount of rivet head deterioration were also evident
on both the skin and apron plates.
A painting contract is currently under way to restore the exterior of the
roller gates and replace deteriorated and worn seals and seal plates.
Summary of Findings
In general, all of the gates inspected were operable and showed no
significant structural distress. However, corrosion was observed during
each gate inspection. While significant corrosion damage was observed
on the out-of-service lift gates at Bankhead Lock and Dam, pitting corrosion and/or blistering paint were visible on all in-service gate structures.
31
In addition to the effect of corrosion on structural integrity, the mechanical systems can also be adversely affected. For example, as observed at
Bankhead Lock and Dam, in lift gates the performance reliability of submerged mechanical systems is an important consideration. Through a regularly scheduled painting and maintenance program the effects of
corrosion can be controlled.
32
Task 3- Environmental
Efects on Gates
Chqirw
33
flexural member. Also, localized corrosion of a structural member can reduce its local buckling strength. A loss of fatigue strength also may occur
because of pits and notches resulting from corrosion. These pits and
notches are stress risers. In areas of high cyclic stress, fatigue cracks, oriented perpendicular to the alternating applied tensile stress, may form at
the pits and notches resulting from the corrosion. These cracks will also
cause failure of the paint film and will result in a visible rust line.
In areas of high static tensile stress, a loss of strength also may occur
due to "stress corrosion" (if the structural material is susceptible to stress
corrosion). However, this is unlikely with material such as A7 steel in a
freshwater environment at ambient temperature.
A buildup of corrosion products can be particularly damaging at
connection details. This can lead to extremely high pin friction in a tainter gate trunnion and may ultimately prevent rotation and gate operation.
A similar buildup of corrosion product at the axles of lift gate wheels
could cause those wheels to "freeze" and lead to an excessive hoist load.
At connections between adjacent plates or angles, a buildup of rust could
cause prying in the riveted connections which can add excess tension
force to the rivet and cause loads to be transferred with eccentricity and
unwanted bending at the connections.
In order to prevent long-term structural damage, corrosion must be controlled through a program of inspection, evaluation, and maintenance.
This general conclusion applies for all spillway gates, not just riveted
ones.
34
7ilWlr
,
Corrosion at crevice
'
between angles
WW
Fge9Vtaltaa..
Blistering of paint
paintsystemfailurne
bt
Am
g
i
Figure 19. Vertical lift gate at Bankhead Lock and Dam with paint blistering and
corrosion
Corrosion
at welds
evident from
paint
sytmfailure",
Il
S..
,/
.'
''
.":paint
;. " ,:
Corrosion adjacent
to rivets evident from
system failure
ioi
Figure 20. Tainter gate at Mississippi Lock and Dam No. 5 with seal weld and
localized corrosion
35
Failure of
top of rivetpaint
due on
to
Failure of paint
at edge
ofagedue
to poor
paint
of paint at edge
()angle due poor paint
system on edge of stiffener ,:
oFailure
a.oen
Sto
rivet heads
croonaedeadcroinon
cosemrrosion.aio
ra paitmosph teri
Failure of paint du
debris accumulation
Figure 21. Vertica lift gate at Bankhead Lock arid Dam with crevice corrosion,
corrosion at edges, and corrosion on rivet heads
a. General atmospheric corrosion.
b. Localized corrosion.
(1)
Crevice corrosion.
Galvanic corrosion.
Filiform corrosion.
Erosion corrosion.
36
Fretting corrosion
Gum
olm
eorrosion
General atmospheric corrosion is defined as corrosive attack which results in slow, relatively uniform thinning. It is expected to occur in the
ambient environment of spillway gate structures, but is not likely to cause
significant structural degradation since the corrosion is spread over a wide
area.
Localized corrosion
SLocalized corrosion is the type of corrosion most likely to affect riveted gate structures. And, because it occurs at specific sites and with
faster rates than general corrosion, it warrants more concern. All five of
the types of localized corrosion are possible on gate structures.
Crevice corrosion. Crevice corrosion occurs in narrow openings between two contact surfaces, a condition prevalent with riveted gates, making crevice corrosion a strong concern. Typically, crevices occur between
adjoining plates or angles (Figures 19 and 21), or between a rivet head
and the adjoining plate or angle (Figure 19). It can also occur between a
steel component and a nonmetal one (under the seals, a paint layer, or debris, sand or silt, or biological organisms caught on the gate members). It
can lead to blistering and failure of the paint system.
Pitting corrosion. Pitting corrosion occurs on bare metal surfaces as
well as under paint films; it is characterized by small cavities penetrating
into the surface with little extension along the surface. If pitting occurs
under paint, it can result in the formation of a blister and failure of the
paint system. Although closeup photographs were not obtained, pitting
was observed on the roller gate that was undergoing repainting during the
onsite inspection at Lock and Dam No. 9 (Figure 6).'
Galvanic corrosion. Galvanic corrosion can occur in gate structures if
steels with different electrochemical potential are used to construct or repair these structures. Generally, when a structure contains mixed steels,
the more electrochemically active steel should be the one having the most
surface area because it will be the steel exhibiting corrosion. This means
that, to avoid galvanic corrosion, rivets and other fasteners with small surface area should be selected to be less electrochemically active than the
structural steel plates or angles they connect. Galvanic corrosion is evidenced by blistering or discoloration of the paint and failure of the paint
system adjacent to the contact area of the two steels. The corrosion at the
rivets in Figure 20, identified above as crevice corrosion, could possibly
also be galvanic corrosion. Galvanic corrosion decreases as the distance
from the metal junction increases.
M. R. Kaczinki. repoa of trip to SL Paul District Office, nmeo to project file, June 17,1992.
Antacnent to letter, J. E. Bower to C. Chasten (WES), June 17, 1992.
Effects on Gates
37
38
attack.
b. Relative humidity. Corrosion of steel is significantly reduced when
the relative humidity is less than 40 percent.
c. Time of wetness. The longer the time of wetness, the greater the corrosion. Importantly, though, corrosion is aggravated by alternately
Effects on Gates
39
Tack weld
Diverter plate
Figure 22. Failure of paint system on tainter gate at edge of plate near tack weld,
Mississippi Lock and Dam No. 9
- .edge
"
,',,m=log.
Corrosion at
of ho le
Corrosion on
to p o f rive t
Z- : ,
Figure 23. Vertical lift gate at Bankhead Lock and Dam with corrosion of rivet
heads
41
42
As described in Chapters 2 and 3, previous Corps experience and current observations during the site reviews indicated that Corps spillway
gates (particularly tainter gates) are subject to occasional flow-induced
vibrations, which cause cyclic stresses in the gates.
However, spillway gates have historically been designed assuming that
all structural loads are applied statically (USACE 1966, 1962). Therefore,
this chapter considers the possible causes of repeated or cyclic loads on
riveted spillway gates and the locations and fatigue strength of riveted details on these gates. In addition, the fatigue-related effects of adding
welds to the riveted gates are discussed.
Gate lifting
During the routine operation of lifting spillway gates, cyclic loads are
applied to structural members from two sources. The first source of cyclic loads is the change in hydrostatic pressure on a gate as it is pulled out
43
of the water and then resubmerged. Although this load case has the potential to produce large variation of stress in structural components, the frequency of occurrence (a worst case assumption is one cycle per day) may
be too low to cause fatigue damage. The other potential source of cyclic
loads is the damped vibration of impulse loads required to overcome the
friction at side seals, particularly in heavy ice conditions and at trunnion
pins.
The possibility of repeated loads due to gate lifting operations causing
fatigue damage is unlikely. However, the potential for damage to structural components due to a single gate lift in extreme (loading) conditions
is a serious consideration.
Flow-induced vibration
This phenomenon produces the most significant cyclic loads on spillway gates because of the potential to combine live load stresses above the
fatigue damage threshold with high frequencies. In fact, any gate which is
discharging water may have some level of flow-induced vibration. Past
experiences at spillway gates along the Arkansas (USACE 1971) and the
Upper Mississippi Rivers (USACE 1986) have indicated that tainter gates
are susceptible to severe vibration problems. From investigative studies
at these sites, it has been found that the problem seems to be heavily influenced by flow conditions (i.e., gate opening and tailwater elevation) and
bottom seal details. Vibrations were successfully reduced or eliminated
by modifying the seal detail or simply operating the gate outside of the
range which causes vibration.
44
100
N - Number of Cycles
Figure 24. Current fatigue strength (Sr - N)curves of AASHTO and AREA
For riveted structural members, however, no definite Sr - N curve has
been established for fatigue design and evaluation. Current design provisions of the American Association of State Highway and Transportation Officials (AASHTO) (AASHTO 1989) and the American Railway Engineers
Association (AREA) (AREA 1992) specify that the fatigue strength of riveted members be determined using the fatigue strength (S. - N) curves for
Category C and Category D, even though the curves were developed for
welded details. According to current AREA standards, the Category D
curve should be used for riveted members when Sr is 12 ksi or greater, and
Category C'should be used when S is less than 12 ksi. It was concluded
in Fisher et al. (1987) that the Category D curve provides an adequate reference for fatigue crack detection in riveted members, and the Category C
curve is a reasonable estimate for fatigue strength before failure.
More test data are needed for the determination of a definite Sr - N
curve or a set of curves for riveted members and connections, especially
in the low-stress-range, high-cycle region. Research on this subject and
on the fatigue crack growth rate in riveted members is currently being conducted by the ATLSS Center.
45
on
Crack origin
Crack origin
Figure 27. Fatigue crack from corrosion notch into rivet hole
made of all the published data from fatigue testing full-size riveted mem-
bers (Fisher et al. 1987). The database expands that of Fisher et al. (1987)
and covers test results from as early as 1937 to date. The test members included full-size riveted connections under tension and full-size riveted
built-up members under bending. The available data from these fatigue
Chapter 5 Task 4 - Effect of Repeated Loading on Gates
47
tests are not abundant and are plotted in Figure 28 with the AASHTO fatigue strength (S - N) curves of Categories C and D, which have been developed for welied details.
0 0|,U
0101
N -Nunteof Cycles
Figure 28. Available Waigue testing data from fu-size riveted members
Although the available test data are not abundant and the conditions of
failure at termination of testing are not defined, some conclusions can be
drawn from these data. As seen in Figure 28, Category D provides a reasonable estimate of fatigue strength for structural details in full-size riveted members subjected to stress ranges higher than 10 ksi (St 2 10 ksi),
while Category C is a lower bound for the lower-stress-range, high-cycle
region. There are insufficient data for a conclusion about the fatigue limit
of riveted members; but no fatigue failure has ever occurred when the
stress range (S) was below 6 ksi (Fisher et al. 1987), provided that the
structure member or detail is not otherwise damaged or severely corroded.
Variable-amplitude fatigue loading
Most of the fatigue test data and the Sr - N curves in Figure 24 were
established from constant-amplitude cyclic loads. In reality, however,
structural members are subjected to variable-amplitude cyclic loads resulting in a spectrum of stress ranges. Figure 29 gives an example of a stress
range histogram compiled from measured live load stresses at a structural
detail. In this figure, the abscissa is the stress range, S , and the ordinate
represents the fractional frequency-of-occurrence, cz. l&or example, the
and S, in Figure 29 show that 6.5 percent of the stress ranges are between
3.75 and 4.5 ksi.
0.50
0.40
0.30
Sm2.9kof
0.20
..
1.5
S ,.- 8.75 ks
so
3.0
4.5
8.0
7.5
0ram
Y.a
(1)
49
lamV
101
10
N Nmnebrof Cycle
-ode
Fgu 30. Reomnde S-Ncuvelfr
PWA.
Tno criteria described above are for riveted spillway gate members
with only minor deterioration. For severely corroded members or members with corroded, loose or missing rivets where the clamping force is reduced or lost, lower fatigue strength curves may be more appropriate.
Specifically, it is suggested that the Category E or E' curves and the conresponding fatigue limits should be used, depending on the degree of corrosion. This is because, as shown earlier, fatigue cracks may initiate at
corrosion notches instead of from rivet holes.
Figure 31 demonstrates the procedure of estimating the fatigue life, N,,
of a riveted spillway gate member which is subjected to variable-amplitude stress ranges represented by Figure 29. In this cme, the highest
stress range, SI, in the stress-range histogram is greater than 6 ksi, the assumed fatigue imit; eventual fatigue damage may occur. The estimated
fatigue life is obtained by locating the intersection of the equivalent stress
range, S , and the extension of the Category C line, as shown in Figure 31
(Keating and Fisher 1989; Fisher et al. 1987).
The recommended fatigue strength criteria for riveted spillway gates
can be summarized as the following, for undamaged and noncorroded riveted spillway gate members and details:
When S,,: 6 ksi, the possibility of fatigue damage can be ignored.
When S < 10 ksi, use Category C and S to characterize the fatigue
strength and life of the riveted member detaJi.
When S > 10 ksi, use Category D and S to characterize the fatigue
strength and life of the riveted member detail.
50
10000
10,
10
-a
1O
Ids
N - Number of Cycles
Figure 31. EsUtimaon of fatgue Ote of riveW spway gates members
At the detail to be evaluated, S, and ,, are the highest stress ranges
and the equivalent constant-amplitude stress range, respectively.
Loading on Gds
51
52
Dam No. 6 in the St. Paul District.' Figure 32b indicates cracks reinitiating from previous repair welds. In this instance, attempts to strengthen a
riveted gate by adding welded stiffening plates creawu very poor fatigue
details. Repairs such as this should be avoided. Figure 33 shows a fatigue crack starting from a tack weld on a riveted bridge member. The
crack initiated at the toe of the tack weld and grew into the riveted plate in
the direction perpendicular to the primary tensile stress. Similar damage
could occur on a riveted gate member.
Figure 33. Fatigue crack Initiating from a tack weld on a riveted member
53
Task 5 - Structural
Evaluation Guidelines
54
Preinspection assessment
Preinspection assessment consists of reviewing design drawings, previous evaluation reports, and all operations/maintenance records since the
most recent inspection. In addition, all critical areas or components of the
gate structure should be identified.
Before travelling to the site and conducting an inspection of the spillway gates, the inspector should prepare by reviewing all available documentation. In the review of existing documents, the following question
should be asked: "Are there suspected preinspection conditions, or have
critical circumstances occurred since the most recent inspection?" These
conditions include the following:
a. A history of problems.
b. Newly reported extreme loads.
c. A change in operational practice.
d. Unusual events.
Specifically, to answer this question, the inspector should do the
following:
a. Review the structural drawing of the gates to become familiar with
the gate components anm operation. Locations and details on the riveted gate structure subjected to high stresses or prone to fatigue
damage and susceptible to corrosion should be identified. These locations should receive more attention during the inspection.
b. Review previous inspection reports, if any, to find out whether any
structural damage or potential problems were noted on the last inspection. If problems were detected, make note of the location and
determine if any repairs were performed. In the case of fatigue
cracks repaired by welding, the cracks often reinitiate at the same locations if no changes to the structural configuration have been made.
c. Search for notes of special events, such as records of gate vibration,
55
Structural elements. For simplicity, the identification of critical structural locations on riveted tainter and vertical lift spillway gates can be subdivided into the following areas: lifting assembly, main framing
members, and skin plate. Because of the unique method of operation of
roller gates, the identification of critical locations is better subdivided into
only two areas, main framing members and skin plate. A more comprehensive discussion of all components of gate structures was presented
in Chapter 2.
a. Tainter gates. The critical components in the lifting assembly on a
tainter gate consist of the hoist mechanism, trunnion assembly, and
seals. A nonredundant chain or wire rope located at each end of the
structure is usually used to lift spillway gates. Because these member are nonredundant, they should be kept in good condition with little corrosion. The trunnion assembly should be kept well lubricated
to prevent excessive friction forces from being developed when the
gate is being lifted. Side seal rubbing plates should be kept clean
and smooth to prevent corrosion and should be free of ice before attempting to lift a gate in extreme weather conditions.
Main structural framing members which should always be carefully
inspected include the vertical ribs, horizontal girders, and end
frames. Because of the application of concentrated loads at the
hoist bearing plate, ribs in this vicinity should be carefully examined for local buckling. Both the horizontal girders and the end
frames should be inspected for global and local buckling (particularly near the knee brace intersection) and should have properly located and unclogged drain holes to prevent corrosion.
The skin plate should be inspected for corrosion loss, missing or deteriorated rivets, and damage due to impact from debris. Although
it is difficult to visually inspect the entire surface of the skin plate,
most problems would be expected to occur in the upper section of
the gate in the splash zone.
b. Lift gates. The critical components in the lifting assembly on a vertical lift gate consist of the lifting hooks, end bearings, tracks, and
guides. Because of their nonredundant application and the concentration of loads at this location, the lifting cable and hooks should
be well maintained. The end bearing assembly should be properly
lubricated and inspected to ensure that the wheel alignment and
track surface finish are at tolerances which prevent local overloads.
To minimize lifting loads, tracks should be flat, and both tracks and
guides should be free of debris and corrosion.
The main structural framing members in a lift gate consist of the
horizontal girders and end posts. Because of their horizontal orientation, the webs of the girders should be inspected for corrosion and
the drain holes should be kept clear of debris. The top girder should
56
57
materials.
Fatigue sensitive details. A checklist of locations (both specific and
general) which are susceptible to fatigue cracking and fracture is presented below to assist the inspector during the preinspection. While reviewing plans of the spillway gate, the inspector should document any
fatigue-critical areas located on the structure. Locations include:
a. Previous cracks repaired by welding, either by deposit of weld beads
over the cracks or by covering the cracks with doubler plates. Figure 34 shows an example of cracks redeveloped at weld repairs.
b. Locations where the structural configuration is similar to where
cracks from rivet holes have been detected.
c. Riveted connections between components such as that between a
roller drum cylinder and the end shields (Figure 35) where the rigidity of the connection prevents the movement of one component
against the other. When a gate is being opened or closed, or when
high-velocity water flows by the gate, relative local displacement
may occur between two rigidly connected components and induce
high stress ranges (live load stress). If the occurrence of these
stresses is frequent (high number of cycles), fatigue cracks may
develop.
d. Flange-to-web junction of horizontal girders opposite skin plate ribs,
where concentrated loads are transmitted between the ribs and the
girder. Localized concentrated loads sometimes induce high local
stresses and movement of the connection angles, resulting in fatigue
cracks from rivet holes in the connection angles.
e. Seal welds in skin plates, particularly when the seal weld is subjected
to repeated loads during operation of the gate.
I.
58
li..
Figure 34. Fatigue cracks redeveloping at weld repair sites on roller gate end
shield
Figure 35. Fatigue crack at connection of roller gate cylinder and end shield
Chapter 6
59
attachments such as stiffeners or repairs which were added to riveted members are fatigue-sensitive areas.
h. All stress concentration areas such as the hoist cable connection to
skin plates.
i. Areas where corrosion has reduced the cross-sectional area and created notches in members.
Inspectlon
The activity of examining, measuring, testing, gauging, and using other
procedures to ascertain quality, detect defects or deterioration, or otherwise appraise a gate and its materials, components, and systems should be
done while the gate is in use and, to the extent possible, raised out of
water. To effectively conduct an inspection, the gate structure should be
systematically examined from one end to the other and from the top to the
bottom. Particular attention should be given to the critical locations cited
in the preinspection assessment. All observations should be documented
in sufficient detail.
While inspecting the various components of a spillway gate, the following questions should be asked:
a. Is there a condition to report for the main structural elements?
(1) Overall or local buckling or deformation.
(2) Loose rivets.
(3) Fabrication defects.
b. Is there a condition to report for the mechanical/electrical
components?
(1) Seal problems.
(2) Hoist guides, bearings, heaters.
c. Is corrosion evident?
(1) Paint loss.
(2) Cross section loss.
(3) Discoloration.
d. Are any fatigue cracks visible?
60
At hoists.
g. At seals.
h. At electrical connections.
61
62
thickness variation are available but are probably not economically justifiable for insitu use on gates.
Radiography, or X-raying, is an NDE technique which can be used to
determine thickness loss due to corrosion, but it only has an accuracy
of 2 percent of the wall thickness and thus is useful only for loss
greater than this. Moreover, the technique requires access to both sides
of the part being radiographed, which adds to the time and cost of the
technique.
Newer methods of inspecting for corrosion are developing, such as
magnetic resonance testing, but these are not yet ready for routine
implementation.
Inspecting for fatigue cracks. A recommended procedure for inspecting riveted spillway gates for fatigue cracks is presented below. This procedure can be followed at all fatigue-sensitive areas on the structure
identified during the preinspection assessment.
a. Examine the gate. Visual examination, particularly with the aid of a
magnifying glass (5x or higher), is the most efficient first step.
b. Confirm presence of crack(s). If cracks are suspected and the gate
component is dry, liquid penetrant can be used to confirm the presence of a crack. More sophisticated methods, such as the use of ultrasonic and magnetic instruments, can also be employed but may
not be needed.
c. Record the location, orientation, and length of the cracks. Record
conditions of the gate when cracks are detected.
d. Take photographs of all cracks showing their position relative to the
components of the gate structure.
e. Compare the conditions of the detected cracks with previously located cracks, if any, before last repair.
Inspecting for deteriorating rivets. As observed during the onsite inspections, deterioration of rivet heads due to corrosion does occur. This
deterioration can be critical and must be looked for during inspection.
The consequences of rivet-head deterioration are: the rivet can no longer
sustain the applied tensile force due to smaller head area; the rivet becomes loose and can no longer hold the connection tight; joint behavior
such as prying may be exacerbated due to reduced head projection beyond
the shank of the rivet; and the rivet will need to be replaced because it is
missing or one of its heads has excessively deteriorated.
Figure 23 shows where rivet heads have split; that is, the rivets have developed rosette heads. These are readily observable in an inspection of a
gate. Headless rivets are also easy to observe, if deterioration has
C tpTr
63
progressed that far. Generally, the degree of head loss (the degree of reduced projection from the shank) should be noted. Is the projection 20
percent, 50 percent, or 70 percent of the original head projection?
In some cases, a corrosion pattern around the rivet will suggest looseness or further corrosion occurring somewhere beneath the rivet head.
Figure 20 shows such a corrosion pattern. The pattern may result from
crevice corrosion where corrosion has penetrated the crevice between the
rivet head and the connected parts; it may also result from a crack that has
formed at the edge of the rivet hole; or it may result from looseness and
motion between the rivet head and the connected parts. The corrosion pattern should always be recorded in these instances.
Inspecting for loose rivets may not be possible using only visual
means. Supplemental inspection tests should be done if loose rivets are
suspected. A commonly practiced nonvisual inspection technique is to
transversely impact the rivet head with a hammer and judge the "give" or
"ring" of the head. A newer technique is to impact the rivet longitudinally
with a commercially available, instrumented impact hammer. These hammers, which generally include a built-in load cell, provide a vibration signal from the load cell. By capturing this signal with a monitoring unit, the
signal can be compared with the vibration signal emanating from a rivet
known to be tight or to the vibration signal resulting from tapping a separate, tightly clamped connection. The magnitude of the impact force must
be consistent through these comparisons. Generally, the signal from a
loose rivet will have a lower and broader frequency content than the signal from a sound rivet.
Awssement
Assessment is the systematic collection and analysis of (a) inspection
data, (b) documents including drawings and previous assessments and
evaluations, and (c) loading and performance criteria regarding an existing gate, which relate to the continuing normal use of the gate.
What will happen if the condition continues?
a. The condition is minor, and of no consequence to normal use.
b. There is some deterioration or problem, but operation and structural
capacity are not jeopardized at this time.
c. There is advanced deterioration or a serious problem and either the
operation or the structural capacity could be affected.
Guidelines are presented below on assessing six conditions commonly
encountered in the evaluation of riveted spillway gates: buckled or plastically deformed structural members, damaged lifting mechanisms, corrosion, fatigue cracking, rivet replacement and weld repairs.
64
65
the members.
Corrosion typically is a slow-growth process. Therefore, if rapidly developing corrosion is reported, air and water samples should be collected
and analyzed to check for unusual changes. If crevice corrosion is becoming more severe, consideration should be given to maintenance of the region during a regular maintenance period, including eliminating any
buildup of corrosion product and then recoating with a more effective
crevice-sealing coating or paint.
When specific members are affected by corrosion, the portion of the
member most affected should be assessed. For example, if the member is
a flexural member, does the corrosion have an effect on either the bending
moment capacity or the shear capacity of the member? Measuring the remaining section thicknesses may be necessary to make this determination.
If the section properties are reduced more than about 10 percent. a new
analysis of the structure should probably be made. If the reduction is
greater than 25 percent, consideration should be given to replacing the
member during the next scheduled maintenance if the member is functionally critical.
Members with fatigue cracks. When a fatigue crack is detected in a
gate structural component, the important items for consideration in assessment are the following:
a. Which member and detail exhibit the crack?
b. Is the crack growing?
c. Is the crack significant if it continues to grow?
d. Is remedial action needed urgently?
Cracks in nonredundant members or components are significant and require more attention. Hoist cables and attachment points of lifting mechanisms to the gate, for example, are nonredundant; continued crack growth
could cause fracture of these components and impair the operation of the
gate. Cracks such as those in the roller gate end shield shown in Figure 32 do not pose danger of imminent failure of the gate.
To determine whether a crack is growing and what the growth rate is,
requires nondestructive monitoring. However, unless the crack is caused
by flow-induced vibration and is in a nonredundant component of the
gate, nondestructive evaluation usually is not necessary.
66
67
Recommendations
The final step in a structural evaluation is the recommendations. The
recommendation is the process of determining the structural adequacy of a
gate for its intended use and making fragment judgments about remediation of problems and frequency of inspections. Personal and subjective
judgment by the persons functioning as expert evaluators is implied. The
question to be asked is: What needs to be done to remedy the condition?
Possible answers include:
a. Continue normal gate operation with only a periodic watch on a specified condition.
b. Continue normal gate operation with a programmed outage for either
repairs, another inspection, or performance measurements.
c. Alleviate the loading by changing gate operational practice, before
continuing gate use.
d. Discontinue gate use and make urgent repairs and/or measurements.
The current Corps inspection and evaluation program specifies that
spillway gates must be inspected every 5 years (USACE 1988). This inspection interval is adequate if a thorough evaluation of the gate structure
reveals no evidence of distress or potential failure. However, if significant deficient conditions exist (e.g., heavy corrosion, fatigue cracks, or deformations) or severe operations occur (e.g., persistent vibrations), it is
recommended that a shorter inspection interval be used to ensure the structural and operational integrity of the gate structure. Because the conditions at each site are unique, proposing a general guideline for selecting
shorter inspection intervals would be difficult and should be evaluated on
a case-by-case basis. An example of selecting a reduced inspection interval is presented in the application examples which follow.
Maintenance operations should be continuous since a comprehensive
program can reduce the occurrence of significant structural distress. In
particular, through a regularly scheduled cleaning and painting program,
the effects of corrosion can be controlled; and by removing debris and lubricating all mechanical components, the potential overloads from lifting
operations can be minimized.
68
Application Examples
EXapl* 1
An example is presented below to demonstrate the proposed guidelines
for conducting a structural evaluation of riveted spillway gates. This case
study is based on the results of the May 28, 1992, inspection of the riveted
tainter gates at Lock and Dam No. 5 on the Upper Mississippi River near
Winona, MN.'
Following the example application, which is based on inspection resuits, an additional example is presented which is based on hypothetical inspection results. This additional example assumes that significant cyclic
stresses have been measured in the gate, and an assessment must be made
of the structural integrity.
Preinspectlon assesment. The tainter gates at Lock and Dam No. 5
are 35 ft wide, 15 ft high, and 25 ft in radius from the trunnion pin to the
face of the skin plate. The structure is framed similar to the design and detail provisions for tainter gates in EM 1110-2-2702 (USACE 1966) with a
3/8-in. skin plate, C12 x 25 vertical ribs, two W30 x I18 horizontal girders, and W18 x 80 strut arm frames. All connections are riveted except
for the use of bolts at the strut arm-trunnion block detail. The non-submersible gates use Type J side and bottom seal details, which have been vibration-prone in the past.
The gates at Lock and Dam No. 5 have a history of structural problems
which include significant gate vibrations (USAED, St. Paul 1988) and
buckled web and flange plates on the strut arm. 2 No extreme loads or unusual events were reported in the time interval since the last inspection. A
change in operational practice was instituted to avoid gate opening settings which cause structural vibration. Because of the history of problems
at this site, a thorough visual inspection was made previously on several
gates.
Inspection. On May 28, 1992, visual inspection was made of all 28 riveted tainter gates at Dam No. 5, and a more in-depth inspection was made
of gates 23 and 24. Weather conditions at the damsite during the inspection were sunny and warm with temperatures in the 70's. The in-depth examination was conducted while water was being released from the gates
and followed the inspection checklist provided earlier in this chapter.
The results of the examination are listed below:
I M. R. Kazkl, report of trip to SL Paul District office, memo to project file, June 17, 1992.
Attached to letter, J. D
fower to C. Chasten (WES), June 17,1992.
2 K. Hobens, Tainter gate strut arm bending at Lock md Dmn No. 5, memorandum for record
(with SL Paul District computation sheets), July 19,1989.
69
f. Corrosion. Paint loss and blistering were visible along the top surface of the web on the upper horizontal girder under the diversion
platc. Blistered paint was left intact during the inspection.
g. Fatiguecracking. No fatigue cracks were observed.
h. Vibration or other unusual behavior. To check for vibration, a tesm
was conducted with the aid of a lockman. Gate No. 23 was fully
closed and then reopened approximately 0.1 ft when vibration began
normal to the face of the gate. By rough measurement, the vibration
frequency was estimated at 5 to 10 Hz. The amplitude of vibration
was maximum at midspan of the gate and was sufficient to create an
audible noise and make ripples in the backwater. The vibration
ceased when the gate was opened further.
i. Application of unusual lcads. No unusual or extreme loads were
reported. There was, however, an extensive accumulation of debris
on the structural memb -- in back of the skin plate, primarily large
timber pieces.
Assessment. Because several detrimental conditions were detected during the inspection, the structural integrity of the spillway gate under normal operating conditions must be assessed. The assessments include the
following:
a. Since the amount of local buckling on the strut arms has not increased since the last inspection and no sign of global buckling was
observed, it is believed that neither the structural capacity of the
buckled members nor the structural capacity of the gate is in jeopardy at this time.
70
b. The amount of water leakage from the side seals is considered tolerable and will have no effect on normal gate operations.
c. Misalignment of the chain hoist is not severe enough to jeopardize operation of gate 25, but should be corrected.
d. Deterioration due to corrosion and rivet bead loss are considered
minor and will have no effect on normal gate operations or gate
strength.
e. Although flow-induced structural vibrations can cause serious damage to the spillway gate, previous field studies have calculated
stress ranges of approximately 4 ksi (USACE 1986). Although this
stress range is below the 6-ksi threshold for fatigue crack growth at
riveted details, the presence of groove welds to water-seal gaps between adjacent skin plates and tack welds to attach the diversion
plate to the gate ribs may reduce this threshold stress range. However, since no fatigue cracks were detected and it is known how to
control the gate vibrations, the structural capacity is not in jeopardy.
71
Example 2
Let it be assumed that during the inspection of tainter gates at Lock and
Dam No. 5 it was reported that a more significant mode of vibration had
recently been observed by Corps personnel. Because of this new information, a thorough inspection was made at all fatigue-sensitive details on several gates where this vibration was observed. However, no fatigue cracks
were visible.
Based on the inspection findings in this assumed example, a field study
was recommended to determine the significance of these new vibrations.
The results of the field study revealed that vibrations of approximately
5 cps (Hz) were producing cyclic stresses of up to 8 ksi at several fatigue
sensitive details on the riveted structure.
The integrity of the riveted gate structure must be assessed by determining the fatigue strength of the details which are subjected to these cyclic loads. To evaluate the remaining fatigue life of the members, the
procedures recommended in Chapter 5 will be followed. Since the measured maximum stress range is less than 10 ksi, the Category C curve will
be used to determine the approximate number of cycles to failure at the detail (this does not imply that the entire structure will fail). By projecting
lines on the Sr - N Curve shown in Figure 31, it can be determined that the
number of cycles to failure is approximately 12.5 million. With the measured frequency of vibration equal to 5 Hz, it would take approximately
694 hr (29 days) of vibration at this stress range to exceed the fatigue
strength of the riveted connection. But because this new mode of vibration has only recently been observed, it is probable that not many cycles
have accumulated to date. In fact, unless the gates in this assumed example are allowed to vibrate for extended periods, it may take up to 3-1/2
years before fatigue cracks develop if vibrations are limited to 1/2 hr per
day while the gates are being adjusted.
The recommended action to address this assumed condition would consist of the following hree steps:
a. Minimize the occurrence of gate vibrations by operating outside the
range which causes vibration.
b. Reduce the inspection interval to approximately I year and inspect a
greater number of gates to ensure that any fatigue cracks will be detected early.
c. Begin engineering studies to determine solutions which reduce the
stresses caused by these vibrations.
72
73
References
74
Refren
75
__....
76
TYP A&DAE"18E
Wag-O
l mmmDlC
lorl
20W&llmIhilItlqml
f *WIl Pa DLarge
074-O
nm.
&d9Kml
Powm
000dinMMIPM4w
Lehigw UWDnity
mysems,
cturalS
1Stu,
f
entIeI frAdvance111dT cnology for
N-1E
Ot.:K
R"
-A-.
N,
lMlll
, PA
180154729O~s.SIVAsS
mdl
1. ASTRTH
(
m)
John E. Bowle, Mark R. Kaczinski, Zouzrang Ma, Yi Zeou, John D. Wood,
Beim T. Yen
&. psRFoRNmORGABNIZTON
9. SPNONUWOIGAGENC
REPORT NUMBER
University,
10.
U.S. Army Engineer Waterways Experiment Station, 3909 Halls Ferry Road,
Vicksburg, MS 39180-6199
PNOAANITMI
N ie aEre
GudO
AGENCY
Technical Report REMR-CoS-43
M1TFWKT)NIAVAILABIUTTY STATEMENiT
Approved for public release; distribution is unlimited.
14L 8UBJECTTERMS
Corrosion
Evaluation
Fatigue1.
Inspection
Riveted spillway gates
MECD
K OO
17. SECURITY CLASMiFICATION I& SECURITY CLASSIFICATION 19. SECURITY CLASSIFICATION 20. UITATION OF ABSTRACT
OF REPORT
UNCLASSIFED
NSNST4O-01-2904=O-
OF THiS PAGE
UNCLASSIFIED
OF ABSTRACT
Standard Form 298 (Rev. 2-89)
Pimel by ANSNIK. S.iS
290-102